Participation in 5 of 10 new BMBF projectsPhysics at Universität Hamburg Develops Digitalization and AI Research

Photo: UHH/MIN/Fuchs

The particle physicist Prof. Dr. Gregor Kasieczka, for example, is developing AI-based simulation processes as part of the KISS collaborative project.

As the only German university, Universität Hamburg is taking part in 5 of a total of 10 collaborative projects being funded by the Federal Ministry of Education and Research as part of the ErUM-Data plan for the next 3 years. The sub-projects at Universität Hamburg will receive a total of €2.1 million. Two of the 5 research collaborations are being coordinated by the Institute of Experimental Physics. They are looking at the use of AI to simulate scientific data and in biomedical imaging.

The collaborative project on AI in rapid simulation of scientific data (KISS for short)

The particle physicist Prof. Dr. Gregor Kasieczka, for example, is coordinating the KISS project in which researchers in particle and astrophysics from all over Germany are taking part. They are developing AI-based simulation processes that, compared to standard methods, facilitate faster, more flexible, and more efficient evaluation of research data. This is necessary because even now particle detectors and telescopes produce large amounts of data for the evaluation of which immense computing resources are necessary. As detectors and telescopes are further developed, calculation capacities reach limits. “Either we will need many more computing resources or we will need much faster and more efficient algorithms to use the extant resources sustainably,” says collaborative project spokesperson Gregor Kasieczka. One of the goals of KISS is to facilitate the more sustainable use of extant resources in data evaluation by developing AI-based algorithms.

As part of KISS, researchers in experimental physics and from the Hamburg Observatory are training algorithms using the data from well-known events until they can precisely and realistically predict events. This research on so-called generative AI methods to evaluate the data of physical processes began in Hamburg with the cluster of excellence Quantum Universe and it can now be developed further in the KISS collaborative project.

Overview of all ErUM-Data collaborative projects in which Universität Hamburg is participating:

• Künstliche Intelligenz zur schnellen Simulation von wissenschaftlichen Daten.
• Coordination: Prof. Dr. Gregor Kasieczka (Institute of Experimental Physics, Universität Hamburg) Partners: Prof. Dr. Marcus Brüggen (Hamburg Observatore, Universität Hamburg), Prof. Dr. Peter Schleper (Institute of Experimental Physics, Universität Hamburg)
• KI-basierte biomedizinische Röntgenfluoreszenz-Bildgebung zum 4D-Tracking von Immunzellen. Coordination: Prof. Dr. Florian Grüner (Institute of Experimental Physics, Universität Hamburg)
• Informationsfeldtheorie für Experimente an Großforschungsanlagen.
• Subproject: Prof. Dr. Marcus Brüggen (Hamburg Observatory, Universität Hamburg) 
• Antizipation 3-dimensionaler Wellenfelder.
• Partners: Prof. Dr. Oliver Gerberding (Institute of Experimental Physics, Universität Hamburg), Prof. Dr. Céline Hadziioannou (Institute of Geophysics, Universität Hamburg), Prof. Dr. Conny Hammer (Institute of Geophysics, Universität Hamburg)
• Optimierte Laserpulse für Freie-Elektronen-Laser.
• Partners: Prof. Dr. Wolfgang Hillert (Institute of Experimental Physics, Universität Hamburg)

The ErUM-Data (website in German) action plan is part of the Erforschung von Universum und Materie program for basic research on research infrastructure (from the smallest structures of matter to the largest structures in the Universe) that the Federal Ministry of Education and Research has been funding since 2017.